Learn Model Context Protocol with Python: Build agentic systems in Python with the new standard for AI capabilities

Christoffer Noring

$31.99 ~~$35.99~~

eBook Oct 2025 304 pages 1st Edition

Christoffer Noring

$31.99 ~~$35.99~~

eBook Oct 2025 304 pages 1st Edition

What do you get with eBook?

Instant access to your Digital eBook purchase

Download this book in EPUB and PDF formats

Access this title in our online reader with advanced features

DRM FREE - Read whenever, wherever and however you want

View table of contents

Preview Book

Learn Model Context Protocol with Python

Explaining the Model Context Protocol

The MCP consists of many parts. In short, JavaScript Object Notation-Remote Procedure Call (JSON-RPC) messages are exchanged between the client and the server. A JSON-RPC message is a message following the JSON-RPC specification, which means it has the jsonrpc, id, method, and params fields, and the data type is JSON. An example could look like so:

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "doSomething",
  "params": {
    "foo": "bar"
  }
}

Quick tip: Enhance your coding experience with the AI Code Explainer and Quick Copy features. Open this book in the next-gen Packt Reader. Click the Copy button

(1) to quickly copy code into your coding environment, or click the Explain button

(2) to get the AI assistant to explain a block of code to you.

The next-gen Packt Reader is included for free with the purchase of this book. Scan...

Transports in MCP

The idea with transports in MCP is that they define how clients and servers communicate. MCP is transport agnostic, meaning it can work over HTTP, WebSockets, STDIO, and more. The transport is the layer that handles the underlying message exchange. It exchanges messages of type JSON-RPC.

MCP has a set of transports it supports from STDIO (for servers running locally) to streamable transports such as WebSockets and Server-Sent-Events (SSEs), and finally request/response transports such as HTTP.

For each of the specified transports, they all have in common that they operate on streams. All the transports have a method exposing those streams, like so:

async with anyio.create_task_group() as tg
    ...
    yield (read_stream, write_stream)

There’s also a BaseSession class that all transports use to send raw JSON-RPC messages, which looks like so.

class BaseSession(
    Generic[
        SendRequestT,
        SendNotificationT,
        SendResultT...

Download Code

Key benefits

The only resource you'll need to build, test, and deploy MCP servers and clients
Take a modern approach toward building, testing, and securing distributed agentic AI apps
Get clear, professional guidance on developing for both LLM and non-LLM clients
Purchase of the print or Kindle book includes a free PDF eBook

Description

Learn Model Context Protocol with Python introduces developers, architects, and AI practitioners to the transformative capabilities of Model Context Protocol (MCP), an emerging protocol designed to standardize, distribute, and scale AI-driven applications. Through the lens of a practical project, the book tackles the modern challenges of resource management, client-server interaction, and deployment at scale. Drawing from Christoffer's expertise as a published author and tutor at the University of Oxford, you’ll explore the components of MCP and how they streamline server and client development. Next, you’ll progress from building robust backends and integrating LLMs into intelligent clients to interacting with servers via tools such as Claude for desktop and Visual Studio Code agents. The chapters help you understand how to describe the capabilities of hosts, clients, and servers, facilitating better interoperability, easier integration, and clearer communication between different components. The book also covers security best practices and building for the cloud, ensuring that you're ready to deploy your MCP-based apps. Each chapter enables you to develop hands-on skills for building and operating MCP-based agentic apps. The Python primer at the end rounds out the practical toolkit, making this book essential for any team building AI-native applications today.

Who is this book for?

This book is ideal for web developers, software architects, AI practitioners, and tech leads looking to build scalable AI-integrated applications using MCP. Product managers overseeing AI initiatives will also find this book valuable. A basic understanding of web development and AI concepts is needed to make the most of this book.